Electronic apparatuses, notably mobile cellular telephones, smart watches and smart glasses, increasingly comprise modules making it possible to carry out a voice recognition function. This function notably makes it possible to control the electronic apparatus vocally.
In order that the apparatus can detect vocal commands at any time, it is desirable to monitor and detect the signals in the environment of the apparatus and notably in its sound environment.
Electronic apparatuses operating on battery power such as mobile cellular telephones are confronted with a problem related to the optimizing of energy consumption during continuous monitoring of the sound environment searching for possible vocal commands.
In order to economize the battery of an electronic apparatus, the latter generally comprises a standby mode in which the electrical power supply is maintained at the minimum in order to keep a minimum of electronic elements operating.
However, when the electronic apparatus is on standby, the latter must come out of its standby state before being able to carry out the process related to the vocal command. A user generally wishes to be able to access the functionalities of the apparatus quickly. It is the therefore necessary to be able to come out of the standby state quickly in order to respond to a vocal command by the user. One possibility is to provide a reactivating button.
Similarly, when the electronic apparatus is not on standby but in an active state, the electronic apparatus does not itself distinguish a vocal command intended for it from a vocal command intended for another apparatus. It is the therefore necessary to provide a method for distinguishing at what time a vocal command is intended for the apparatus.
An example of a known method for detecting the surrounding sound environment is the continuous audio process, or “continuous audio processing”, for waking a telephone. This type of method is known by the expression “Always ON Audio” or “AON Audio”, because the audio voice recognition circuit is kept powered and in continuous operation.
The major disadvantage of such a system is the high energy consumption, given the continuous power supply of the voice recognition circuit.